5 ms·
> However, in that framework, in many cases you're still often left with uniform priors, which then reduce to frequentist statistics. Several commenters have m
by pjcv 7y ago
> However, in that framework, in many cases you're still often left with uniform priors, which then reduce to frequentist statistics.
Several commenters have made this claim, which seems to be true only for maximum a posteriori (MAP) estimates. Frequentist methods do not construct a distribution over parameters.
- skat20phys 7y agoThat's a good point regarding MAP estimates versus, e.g., EAP estimates, although it does lead to questions about whether one or the other is preferable, and I think good arguments can be made either way.